Adjustment for Non-Response Bias in a Rural Mailed Survey

Reasonable reliability of the mailed inquiry has been attained by Agricultural Estimates and other agencies whose restricted budgets require them to rely largely on this method of collecting data for their surveys. Reliability is achieved mainly by building up historical series for some surveys that are desired, different techniques must be used to increase their accuracy. This paper is confined to the statistical analysis of one source of bias in estimates--the bias of non-response--in a survey conducted by the North Carolina Field Office of Agricultural Estimates. It is possible in some instances, according t o the author, to estimate totals of agricultural items reliably by using information from successive waves of mailed inquiries. He is convinced from the accumulated evidence that a general law governing reliability is operating but that further research is needed to learn its precise character.